GitGrilled — Voice Interview Practice and System Design Coaching from Your Codebase

Practice live voice interviews grounded in your actual codebase — with uncompromising interviewers that grill your architecture, implementation tradeoffs, and verbatim source code lines.

One-Word Launch

Replace "github" with "gitgrilled" in any public repository link (e.g. gitgrilled.com/facebook/react) to start a live interview on that codebase instantly — no setup.

Not a Polite Chatbot

Generic voice modes are trained to agree. GitGrilled interviewers give zero praise, and the server enforces real pressure: silence nudges fire at 15, 20, 35, and 55 seconds, and the session ends at 60 seconds of silence. Evade a question and you earn a strike — three strikes end the session early.

The Interviewer Has Read Your Code

During a live session the interviewer searches your repositories, opens files, and checks commit history — and quotes your lines back. Every weakness must cite the exact sentence you said.

Nine Personas, Fixed Rules, No Flattery

The Poker Face gives zero praise ("Understood." "Noted."), the Skeptic challenges the weak link with "Why" pushback capped at twice per topic, the Rapid-Fire Executive demands bottom lines, the Mentor gives Socratic hints, the Purist chases edge cases to the ground — nine enforced behavioral contracts, server-side.

The full loop runs four rounds — Technical & Algorithms, System Design, Codebase & Architecture Defense, Behavioral & Leadership — each with a different interviewer. Prior-round summaries carry forward, so a weak spot from Round 1 gets probed again in Round 3.

Scoring That Can't Be Inflated

Every substantial answer is scored one to five across five dimensions (technical depth, problem solving, system design, communication, code quality). The server computes the weighted overall — the model literally cannot inflate it. Hire bands: 90+ Strong Hire, 75+ Hire, 60+ Lean Hire, 40+ Lean Reject, 20+ Reject, below 20 Strong Reject.

The report includes evidence-linked quotes that jump to the transcript, answer rewrites, alternate paths, a percentile benchmark against your target company preset, a hiring-committee debrief, and a one-click Study Handoff for ChatGPT, Claude, or DeepSeek.

Daily Practice

Quick Drills are weak-spot drills auto-built from the exact gaps your critique found. The Daily Drill builds a streak. The Reference Library holds 41 cheat sheets, frameworks, and example interview transcripts.

Privacy

Code ingestion is zero-retention — files are parsed in-memory and never stored. Raw audio is never saved. GitHub OAuth tokens are AES-GCM-encrypted. Delete any session or purge the whole account from Settings.

Pricing

Free is 15 minutes a day, Pro is 500 minutes a month ($19), MAX is 2,000 minutes a month ($49). Nothing is gated but minutes — repo analysis, scorecards, transcripts, drills, and the library are unlimited on every plan. See full pricing.